The primary objective is to use psychological profiles described by Holtzworth-Munroe & Stuart to identify batterer typologies in a population-based, cross-sectional study of partner violence (PV) perpetration. Three subtypes will be employed: GVA, BD and RO have been described. All participants will be asked about their violence perpetration, and victimization, and their psychological profile. These data will be used to assess batterer subtypes and grouped as (a) current PV perpetrators by batterer subtype, and (b) those with the psychological profile of a batterer yet do not report physical PV (labeled high-risk non-perpetrators). Identification of this latter group will provide an opportunity for a novel application of batterer typologies that may have important implications for prevention research. The investigators will explore protective factors (e.g., no substance abuse, high integration, or strong parental attachment) for PV among this high-risk non-perpetrator group. [unreadable] [unreadable] Aims include estimation of the prevalence of, and risk factors for, both PV perpetration by batterer subtypes and high-risk, non-perpetration among 2000 students sampled from a large community college system in Houston, Texas. Measures of distal and proximal risk factors based on a theoretical model will be used to guide assessment of correlates of both perpetrator and high-risk, non-perpetrators by psychological batterer subtypes over 3 levels of influence (individual, family, and ethnic community). It is anticipated having sufficient numbers of subjects from at least 3 ethnic communities: Mexican-American, African-American, and Caucasian. Point estimates and confidence intervals for PV typologies and perpetration will be estimated by gender. Multinomial logistic regression will be used to assess risk factors (e.g. childhood abuse and witnessing IPV, substance abuse, affect) for the 3 PV typologies by level of influence. To determine whether the novel, high-risk, non-perpetrator group is at increased risk of using force against a partner (AIM 3), a short-term (8-month) cohort study will be conducted to prospectively estimate PV risk among those not reporting perpetration in the cross-sectional interview by psychological profile. All respondents scoring as high-risk, having the psychological profile consistent with a battering subtype yet not perpetrating PV will be included. A similar number of low-risk (no batterer psychological profile and non-perpetrators) will be included as unexposed. Cox proportional hazards modeling will be used to estimate time to first use of physical force by psychological profile. [unreadable] [unreadable]